MQA

[bigshot]

71dB, there is absolutely no way he got anything out of your last reply. You need to first limit yourself to the basics if you want to get across to him how digital audio works. Basic stuff is his level. But he doesn't seem interested in anything anyone else says. He's speaking to entertain himself and we're all just props for his monologue. I think he's one of our regular trolls.

 

[Audiophiliac]

That's a good question for someone getting into the science of digital audio. The simple answer to this is that an ADC doesn't need to "do" anything between the sample points, because the signal to be sampled is properly bandlimited. This means it is mathematically "known" what the signal does between the samples. It can't do anything surprising, because the limited bandwidth doesn't allow that. This is the essence of the sampling theorem behind digital audio.

Well, maybe someone should try to come up with a mathematical formula to play perfect analogue chart-toppers between samples that haven't been written yet.
That's what we should make ai chicks do, so that the chicks who don't have time to be people's ai chicks win. Oh no, some chick will try to program an ai chick that does that for you. Everybody else is going to want to be Decker, man...

Mathematically this is a bit like having a parabola completely defined everywhere by knowing only three points of it. Music is massively more complex "curve" than a parabola and that's why we need 44100 sample points for every second of music, but it is enough the same way knowing just three points of a parabola is enough. Try telling mathematicians you actually need 10 or 100 points of a parabola to define it "more accurately" and they will tell you to study math just as we are telling you to educate yourself about digital audio.

I'm with you about a formula recreating what was going on sounds like sounds good, since you could just sit there all day coming up with formula's with variables describing what things were shaped like. But what variables are we working with? We capture current frequencies playing at the time (khz), at their loudness (bit depth), with nothing describing shapes. But the sound is changing constantly, otherwise you'd be a fool for sitting there listening to garbage. Ever had your audio out lock up, and get the same noise playing till you reset it? That's what tracks sound like if you only use 1 sample. Then you're back to sample points anyhow, and how many is enough.
*** We're just sticking to the theory of the format for now, never mind that the original digital output combined the time signal with the data signal, so if you want a dedicated audio out, you still have to design your own, fortunately general-purpose usb keeps them separate, but they're not all even galvanically isolated, on top of all sounding different. Why is using an hdmi cable what's catching on with dac and transport designers? If you only need 2 wires, 1 for clock, and one for data, wouldn't 2 rca cables have more juicy contact and better cable options than hdmi? Oh no, someone will say they don't care if they paid double every month for mqa gear that sounds better than the gear without it, instead of them just trickling double to us sometimes, because I don't understand it sounds perfect, and they'll say I'm being their enemies. ***

 

[bigshot]

Shapes are defined in time... frequency, amplitude and time are the three factors in sampling. 44.1 can capture the shape of any possible waveform up to 20kHz. It doesn't matter if that waveform defines a test tone or a symphony orchestra. It's perfectly captured up to 20kHz. Above that, there is nothing. It's filtered off because without two points to define a frequency, all you get is noise.

 

[Audiophiliac]

I'm not sure what you mean by that. An ADC indeed does more than just sampling in order the capture information between the samples as accurately as possible. For example, a good ADC will use filtering and dithering to help with properly capturing an analog signal. I'm asking again. Is there anything that could convince you an ADC captures information between its sample points or not?

Filtering means something is being cut out, an adc chip does that to the source?

 

[sander99]

He took another snif/shot/drink/smoke or something just now I think.

 

[Audiophiliac]

shapes are defined in time.

Not if they change.

 

[Audiophiliac]

He took another snif/shot/drink/smoke or something just now I think.

They'll probably tell you it's perfect, and you don't understand the formula.
What if people argue that oversampling makes your chip play what was between the samples even more perfectly? Then we could all oversample to 384, and that's what we want mqa to charge Tidal to do. That's what we like about Master Quality Authenticated, it sounds better than the original. You're the enemy if you disagree.

 

[bigshot]

Filtering means something is being cut out, an adc chip does that to the source?

Frequencies above the range of human hearing are filtered off because 44.1 can't capture frequencies above 20kHz. If you try to run super audible frequencies through, your sample will randomly capture parts of the points of those frequencies, resulting in random noise. So you filter them off to prevent them from adding noise at audible frequencies.

Not if they change.

Samples change 44,100 times a second to capture the shapes of waveforms. That is a huge number of samples. Movies get by with just 24 frames per second. Sound is sampled almost 2,000 times faster.

 

[VNandor]

Filtering means something is being cut out, an adc chip does that to the source?

Technically it's not the chip that does the low-pass filtering, at least not the part that's responsible for the conversion. However if you go into a shop and say you want to buy an ADC, noone will give you just the chip, they give you a kit that let's you capture an analog voltage signal if used properly. If an analog signal contains frequencies above a certain threshold, then the information captured by the ADC between the sample points will be inaccurate. If the incoming analog signal is not band-limited at all, then it would be impossible to capture information between the sample points for an ADC. Because of that, a good ADC will have a specific kind of low-pass filter to band-limit the incoming analog signal. This band-limited signal is what goes into the ADC for the conversion. Keep in mind, any real-world signal (be it changes of sound pressure over time or changes of voltage over time) will be band-limited. If the source for the ADC is a regular microphone that's being used for recording music, then the source will be band-limited because microphone used for recording music can't capture sounds that are significantly above 20kHz. That means if you blast a mic with something that doesn't contain frequencies below ~20kHz, it won't produce any voltage for the ADC to convert. I feel like I'm wasting my time because you keep dodging my question.

 

[71 dB]

Well, maybe someone should try to come up with a mathematical formula to play perfect analogue chart-toppers between samples that haven't been written yet.
That's what we should make ai chicks do, so that the chicks who don't have time to be people's ai chicks win. Oh no, some chick will try to program an ai chick that does that for you. Everybody else is going to want to be Decker, man...

I'm with you about a formula recreating what was going on sounds like sounds good, since you could just sit there all day coming up with formula's with variables describing what things were shaped like. But what variables are we working with? We capture current frequencies playing at the time (khz), at their loudness (bit depth), with nothing describing shapes. But the sound is changing constantly, otherwise you'd be a fool for sitting there listening to garbage. Ever had your audio out lock up, and get the same noise playing till you reset it? That's what tracks sound like if you only use 1 sample. Then you're back to sample points anyhow, and how many is enough.
*** We're just sticking to the theory of the format for now, never mind that the original digital output combined the time signal with the data signal, so if you want a dedicated audio out, you still have to design your own, fortunately general-purpose usb keeps them separate, but they're not all even galvanically isolated, on top of all sounding different. Why is using an hdmi cable what's catching on with dac and transport designers? If you only need 2 wires, 1 for clock, and one for data, wouldn't 2 rca cables have more juicy contact and better cable options than hdmi? Oh no, someone will say they don't care if they paid double every month for mqa gear that sounds better than the gear without it, instead of them just trickling double to us sometimes, because I don't understand it sounds perfect, and they'll say I'm being their enemies. ***

You don't take this seriously. It is pointless to answer to your posts.

 

[Audiophiliac]

Frequencies above the range of human hearing are filtered off because 44.1 can't capture frequencies above 20kHz. If you try to run super audible frequencies through, your sample will randomly capture parts of the points of those frequencies, resulting in random noise. So you filter them off to prevent them from adding noise at audible frequencies.

I thought the argument was that they were unable to be created?

Samples change 44,100 times a second to capture the shapes of waveforms. That is a huge number of samples. Movies get by with just 24 frames per second. Sound is sampled almost 2,000 times faster.

44.1khz is not a huge number of samples, it's the bare minimum required to capture 22 khz. Movies don't get by with 24 frames per second, you hardly get any of it compared to reality. Tell someone to dance behind and to the side of your screen, and see which has smoother motion. Recording motion hasn't been invented yet, only capturing pictures. Strung together pictures are called motion pictures, and that's what videos on your phone of your 2 year old are. Noone has recorded motion yet. But sound in motion was, before digital.

 

[Audiophiliac]

You don't take this seriously. It is pointless to answer to your posts.

I guess that's the problem I'll be facing with an mqa thread, it's only possible to waste your time listening to it.
But please answer my final question. What do you say if someone says oversampling makes your digital gear sound even more perfect?

What if someone wanted to put chips in their ears to hear everything perfectly? If you play your music too loud, you'll probably get into that.
Since they probably don't record it, they get to hear why analogue was better.
Then, since you can't download copies of what they hear, they'll probably tell you some chick in Canada actually went for that crap.

 

[gregorio]

Does that mean digital audio doesn't use samples, and just plays everything perfectly? That's how I would be wrong.

It’s both, digital uses samples and plays perfectly and that’s how you’re wrong but don’t let that stop you.

I don't want to watch a video right now

Of course you don’t, you haven’t wanted to in the past either because you don’t want to know the facts, you just want to keep making up nonsense and epitomising Dunning-Kruger. How long did it take for you to get kicked out of ASR?

G

 

[Audiophiliac]

It’s both, digital uses samples and plays perfectly and that’s how you’re wrong but don’t let that stop you.

Of course you don’t, you haven’t wanted to in the past either because you don’t want to know the facts, you just want to keep making up nonsense and epitomising Dunning-Kruger. How long did it take for you to get kicked out of ASR?

You're avoiding the question. What if someone says oversampling to 352,800khz sounds more perfect than 44.1? Especially since the 44.1 was already perfect, and you're Dunning-Kruger for not getting that.
When 64k video is out, and people watch Citizen Kane on it, it will look like it originally was 64k, except with crappy film and gear. People who thought the video tape copy they made off cable tv and thought it looked like what it should look like will think it looks good, but have already seen it. A couple of times, actually. Yeah, they'll release it. All I know is that he threw his sled into the fire at the end, so he must have lost.

 

[bigshot]

What if someone says oversampling to 352,800khz sounds more perfect than 44.1?

That's easy. You just rack up two identical tracks, one at 352,800kHz and one at 44.1kHz and see if the person can tell the difference in a blind comparison.

But if you want a prediction of how that might turn out, simply looking at the numbers will tell you. 44.1 is capable of reproducing all frequencies perfectly up to the edge of human hearing, 20kHz. There is absolutely no evidence that human beings can hear 176khz. A bat can hear up to 200kHz. That is the only animal on Earth that can hear that high. If you are a bat, you might need a sampling rate that high, but not if you are a human being.